Related Art
Character-recognition systems are widely used to extract data from scanned copies of financial documents, such as checks and vouchers. These character-recognition systems convert printed and/or handwritten text into digital data that can be used by other software, such as accounting systems. While these character-recognition systems have become increasingly accurate over the years, they still make mistakes during the recognition process.
Some character-recognition systems seek to minimize these recognition errors by having a human double-check the recognition process. Typically, a human is presented with the recognized data from the image, and is asked to confirm or correct the data. While this technique helps reduce the number of recognition errors, it also has major drawbacks. First, this process is very time-consuming for the human. (In many cases, the human has to search for the data in the image that he or she is verifying.) Furthermore, in some cases where the number of images is large, the human may become aggravated and may not carefully review the recognized text for errors.